What? What’s a “map claim”?
We have a map that says “X”. If the map also says “X exists”, that’s not adding anything useful or meaningful to the map.
But saying “this is a map of Narnia, which does not exist” ,says something useful.. something which stops you sailing off to find Narnia.
“this is a map of Narnia which should not be used to predict experiences”
Really, using the Solomonoff induction model, you’d just put a very low probability weight on that map since Narnia is complicated to define.
I think you are right that Solomonoff can’t be simultaneously expected to give you both efficient predictions and correspondence to objective reality.
But saying “this is a map of Narnia, which does not exist” ,says something useful.. something which stops you sailing off to find Narnia.
“this is a map of Narnia which should not be used to predict experiences”
Really, using the Solomonoff induction model, you’d just put a very low probability weight on that map since Narnia is complicated to define.
I think you are right that Solomonoff can’t be simultaneously expected to give you both efficient predictions and correspondence to objective reality.